1996 Fiscal Year Final Research Report Summary
A Study on Reliability-based Fatigue-proof Design Method Based on Probabilistic Fracture Mechanics
| Project/Area Number |
07650106
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Materials/Mechanics of materials
|
|---|
| Research Institution | Kagawa University |
|---|
Principal Investigator |
ISHIKAWA Hiroshi Kagawa University, Founding Office of New Engineering Faculty, Professor, 工科系学部・創設準備室, 教授 (60026200)
|
|---|
| Project Period (FY) |
1995 – 1996
|
| Keywords | Reliabilily-based fatigue-proof design / Reliability assessment / Residual life distribution / Probabilistic fracture mechanics / Fatigue crack propagation / Markov approximation / Inspection and maintenance / crack length distribution |
|---|
| Research Abstract |
1. Stochastic models dealing with cumulative fatigue damage process can be categorized into the following six models : (1) random variable model, (2) Markov chain model, (3) Stochastic differential equation model, (4) locally averaged model, (5) renewal process model, and (6) Markov approximation model with the notion of the death point derived by the present study.
Among these models, the last one currently concerned with appears to be the most promising and useful with wide practical applicability.
2. By use of the model, theoretical distributions of crack length and residual life have been derived by taking into considerations a variety of uncertain factors associated with fatigue crack growth process.
3. The derived distributions have been applied to solve various practical problems where safety and reliability play a major role, and their propriety has been exemplified based upon parameter sensitivity studies.
4. Distribution parameters in the models can be classified into two categories : (1) physical parameter appearing in the crack propagation law itself and (2) statistical parameter associated with the randomness in applied loads and material properties. It is clarified that the latter can be estimated even by use of fairly small amount of non-statistical experimental data.
|
